search

VITA-Group / FasterSeg

[ICLR 2020] "FasterSeg: Searching for Faster Real-time Semantic Segmentation" by Wuyang Chen, Xinyu Gong, Xianming Liu, Qian Zhang, Yuan Li, Zhangyang Wang
MIT License
526 stars 107 forks source link

How to use Tensorboard within the FasterSeg project? #38

Closed Gaussianer closed 4 years ago

Gaussianer commented 4 years ago

Hello, FasterSeg is very interesting as part of a research project at our university. We would like to train it with our own data set.

How can I monitor the whole process with Tensorboard? In which directory do I have to be and which command do I have to execute?

We have forked your repository for this and are in the process of providing a guide on how to train FasterSeg on custom data.

We already provide a docker file for the training environment. Other required scripts are currently being implemented.

See https://github.com/Gaussianer/FasterSeg

Best regards

Gaussianer commented 4 years ago

So, now I know - For those who ask themselves the same question, here is the answer for example for the pretraining of the supernet: tensorboard --port 6006 -logdir search-pretrain-256x512_F12.L16_batch3-20200101-012345